Self-tuning Control Of Non-Uniformly Sampled System

Currently the theory research about the identification and control of the uniform sampling system has been quite mature. However, in our daily life and many industrial control processes, as the development of network, communications and so on, some difficulties appeared when the traditional single-rate systems had been used, so multi-rate system has been developed. With variety requirements of industrial control, multi-rate system has been widely applied in many areas. Non-uniform multi-rates system is a special one. Now more and more domestic and foreign scholars are studying on it.This paper proposes a self-tuning control algorithm for the non-uniformly sampling system based on the current researches. Firstly the non-uniform sampling system is decomposed to several subsystems, and then the paper extended apply minimum variance control algorithm to control all the subsystems, so at last the whole system will be controlled. The basic idea is to make the actual output tracking the desired output, using the objective function which is equal to the variance of the actual output minus the desired output, the control amount is the input that can make the objective function minimum.This paper describes three normal type of sampling scheme of the non-uniform multi-rate systems in detail, and build state space model for the two programs. And then it lift variables from the established state-space model. Then the paper use recursive least squares identification algorithm to identify parameters, and at the same time verify the convergence of the identification algorithm. Afterwards the paper analyze the non-uniform sampling system with white noises without delay and containing the minimum phase, and then put forward the corresponding algorithm to control the system and at last verify the feasibility of the algorithm through simulation. Besides, the paper also analyze the non-uniform sampling system with colored noises with delay and containing the non-minimum phase as before Finally the paper design the control scheme for the example of a two-axis milling machine, at the same time take different noise variances to observe the influence of different SNR on identification and controlling. At last this paper summarize itself, describe its deficiency and put forward some ideas for future research.